<p>Another one lost to the siren call of the UNSWR rankings. Too bad.</p>
<p>ModelingLiao, could you do a similar analysis of the last 10 years of Maxim magazine’s annual Hot 100? I’m particularly interested in Jennifer Garner’s relatively short shelf-life compared to Halle Berry’s relatively long shelf-life when it comes to stirring arousal in adolescent males.</p>
<p>It is a well known fact that AA women age better than their white counterparts. See Halle, Tina Turner and many others. ;-)</p>
<p>On post # 20, it should be UCBChemEGrad instead of UCBerkeleyChemEGrad. Sorry I probably keyed in too many Berkeleys on one post.</p>
<p>Quote originally by alam1</p>
<p>umm… Im sure Emory ranked 9th once… i dont think you added that. nice to see Emory has moved up in the rankings over the years.</p>
<p>As discussed in post # 20, in 1989, USNWR introduced new empirical formula, which weighed in other important academic criteria e.g., faulty resources, acceptance rate, class size, etc. and eventually resulted in one of its biggest changes to the order of top 25. After such adjustment, top 25 had never been the same ever since. The following are Emorys ranking summary before, on, and after 1989:</p>
<p>Highest: 9 (1998)
Lowest: 25 (1988, 1994)
Before 1989: 25
On 1989: 22
In the 1990s: 9-25
In the 2000s: 17-20</p>
<p>Quote originally posted by Schmaltz</p>
<p>ModelingLiao, could you do a similar analysis of the last 10 years of Maxim magazine’s annual Hot 100? I’m particularly interested in Jennifer Garner’s relatively short shelf-life compared to Halle Berry’s relatively long shelf-life when it comes to stirring arousal in adolescent males.</p>
<p>I would be glad to do the analysis as requested if I had Maxim magazine’s annual Hot 100 for the last 10 years. Upon searching Maxim magazine’s annual Hot 100 in the internet, I was able to find 2009 list but unable to locate the lists for the previous years. Does anyone know if such information is available via internet?</p>
<p>There are enough horny slobs on this site that I’m sure we’ll be able to find some past Maxim data.</p>
<p>It is a well known fact that AA women age better than their white counterparts. See Halle, Tina Turner and many others.</p>
<p>Yeah, Moms Mabley was hot well into her 70’s.</p>
<p>Quote originally posted by fallenchemist</p>
<p>As far as repititions and duplications, moving up or down a slot or two hardly represents a real change. More like statistical variance within a range. They could do this every five years and get mostly the same results, but given the pace of changes at universities (excepting a natural disaster like Katrina), doing it every year is piling absurd on top of ridiculous.</p>
<p>Fallenchemist: </p>
<p>As I mentioned previously, even though USNews Top 25 did vary year by year but similarities should be expected if same methodology utilized. USNews made many modifications in methodology for model improvement. Such adjustment help USNews better defend their results, which were based on contemporary data collected. You cannot argue with their defensible data, let alone pressing any meaningful challenge on their data interpretation results (top 25). As for the appropriate reporting frequency for college rankings, I would say current annual setting is fine since each year represents each class and each class is a league of their own.</p>
<p>
I very easily can do such a thing, but more to the point I can argue that this represents a list of the “best” colleges. Best for what? Best to whom? Beyond that, the very fact that they “adjust” their parameters, weighting factors, etc. means that the previous lists must have been wrong. Or maybe now they have it wrong. How can you know? They put 25% on peer assessment. How can you assure me that a person at Harvard has any idea what is going on at Drake, or any number of other schools? These people have lives, I don’t think they are spending their time researching other schools in detail. They also have a bias born of location and history. Schools that are more isolated compared to the crowded Northeast and West Coast are at a natural disadvantage in these peer assessments. So yes sir, I can argue very strongly with many aspects of the USNWR rankings. Also:
How did you not contradict yourself from one sentence to the next?</p>
<p>Also, as far as the frequency, you are implying that the stats of a given class make a significant difference in the rankings. In fact their impact is rather small by design, and even more so since schools don’t really change that much in the average stats from year to year, usually. The very fact that, as you say, there are strong similarities in the list year to year shows it isn’t really needed every year.</p>
<p>As a researcher/engineer/modeler, I would always enjoy sharing my ideas/viewpoints/findings/results with others and I often do so via professional conferences and seminars; but all in my familiar areas: Water Quality and Stormwater Modeling. As mentioned in my earlier posts on another thread titled What are the new top25 universities as a senior , this time I simply was struck by the idea of using modeling concept for college ranking application and would like to see such idea explained in a clear-and-open-and-transparent way.</p>
<p>In modeling practices, there are four major phases for developing a model: model design, data collection, data interpretation, and calibration and validation. In Phase 1, modelers perform thorough literature review to come up with an appropriate design. After peer reviews of the proposed model design, it moves to Phase 2, data collection. </p>
<p>Data usually come and/or derive from survey results. In USNews model, subjective matters such as how much weigh should be assigned to and/or distributed among faculty resources/productivity, SAT, % acceptance, class rank, and reputation may vary among different age, sex, ethnic, and demographic groups. Upon getting the feedback from annual survey, those subjective matters are determined from the survey with modelers best professional judgment instead of randomly pick any number set from novices or laymen, which is the completion of Phase 3, data interpretation.</p>
<p>After assigning weigh of the aforementioned subjective matters, the calculated rankings should be compared with those results from previous years to see if there are any big surprises. If yes, based on your best professional judgment, modelers may further examine those surprises to resolve the confounding factors (Phase 4, model calibration or fine tuning). After model is fine-tuned, the model predicted ranking needs to be validated by peer review. (Professionals and/or Readers like you and me) </p>
<p>As years go by, USNews made many modifications in methodology for model improvement. Such adjustment help USNews defend their results, which were based on contemporary data collected. From a modelers viewpoint, you cannot argue with their defensible data and let alone challenge their data interpretation (top 25 results). As for the appropriate reporting frequency for college rankings, I would say current annual setting is fine since each year represents each class and each class is a league of their own.</p>
<p>re-sent for typos^^</p>
<p>As a researcher/engineer/modeler, I would always enjoy sharing my ideas/viewpoints/findings/results with others and I often do so via professional conferences and seminars; but all in my familiar areas: Water Quality and Stormwater Modeling. As mentioned in my earlier posts on another thread titled “What are the new top25 universities as a senior ”, this time I simply was struck by the idea of using modeling concept for college ranking application and would like to see such idea explained in a clear-and-open-and-transparent way.</p>
<p>In modeling practices, there are four major phases for developing a model: model design, data collection, data interpretation, and calibration and validation. In Phase 1, modelers perform thorough literature review to come up with an appropriate design. After peer reviews of the proposed model design, it moves to Phase 2, data collection. </p>
<p>Data usually come and/or derive from survey results. In USNews’ model, subjective matters such as how much weigh should be assigned to and/or distributed among faculty resources/productivity, SAT, % acceptance, class rank, and reputation may vary among different age, sex, ethnic, and demographic groups. Upon getting the feedback from annual survey, those subjective matters are determined from the survey with modelers’ best professional judgment instead of randomly pick any number set from novices or laymen, which is the completion of Phase 3, data interpretation.</p>
<p>After assigning weigh of the aforementioned subjective matters, the calculated rankings should be compared with those results from previous years to see if there are any big surprises. If yes, based on your best professional judgment, modelers may further examine those surprises to resolve the confounding factors (Phase 4, model calibration or fine tuning). After model is fine-tuned, the model predicted ranking needs to be validated by peer review. (Professionals and/or Readers like you and me) </p>
<p>As years go by, USNews made many modifications in methodology for model improvement. Such adjustment help USNews defend their results, which were based on contemporary data collected. From a modeler’s viewpoint, based on their 25+ years of solid modeling efforts, one should not argue with their defensible data or challenge their data interpretation results (top 25). As for the appropriate reporting frequency for college rankings, I would say current annual setting is fine since each year represents each class and each class is a league of their own.</p>
<p>The problem is the entire premise is false. You cannot model something as subjective as a “best” college. You can use all the fancy lingo and methodologies you want, but the fact of the matter is there is nothing scientific about this at all. In the end a model has to be compared to something that exists in the real world, something the model is trying to approximate because the real thing is too complex to analyze otherwise. What is this model trying to approximate? Who is the best? How do you know it is right then? By already knowing who is best? That becomes a circle really fast.</p>
<p>Quoted originally from post #31 by fallenchemist</p>
<p>The problem is the entire premise is false. You cannot model something as subjective as a “best” college.</p>
<p>fallenchemist , youre entitled to your own opinions. As for myself, a modeler/researcher, I believe that a well-calibrated model, under appropriate peer-review guidance, should be able to predict and/or help suggest the best colleges within a pre-defined set of selection criteria (see post # 30). Let us leave it at that and move on.</p>
<p>No, I want to know how you calibrate your model, and what constitutes appropriate peer-review guidance. That is the point. These are just terms to obscure the fact that USNWR is pretending to measure something that cannot be measured, because what is “best” for one person is not “best” for another. Given that (and that certainly cannot be argued with), it is a fool’s errand.</p>
<p>Moms Mabley looked as good at 70 as she did at 35.</p>
<p>Exactly, Barrons, it was like Moms’ looks not only stopped the clock, they stopped the calendar too.</p>
<p>Quoted originally from post #33 by fallenchemist</p>
<p>No, I want to know how you calibrate your model, and what constitutes appropriate peer-review guidance. That is the point. These are just terms to obscure the fact that USNWR is pretending to measure something that cannot be measured, because what is “best” for one person is not “best” for another. Given that (and that certainly cannot be argued with), it is a fool’s errand.</p>
<p>fallenchemist :</p>
<p>It took me many years of graduate school training to have strong grasps of modeling concept and several additional years to master in modeling. I dont think it is a good idea to post very complicated concepts/frameworks e.g., model calibration and validation on this thread. Whereas, you might find in my earlier posts on another thread titled What are the new top25 universities as a senior? which includes a mini-series of top 25 college ranking model development. I started that thread on May 4, 2009 with a mind-set on implementing modeling concept to derive a defensible top 25 ranking so a more detailed modeling approach were included. It is a long thread but I believe it is worth reading it. Let me know if you have trouble finding or reading it.</p>
<p>modelingLiao - why don’t you learn how to use the quote function on this site? Since you are so sophisticated in this kind of thing, I am sure you can figure it out.</p>
<p>I went to that thread and read the first few posts. I can already see you don’t know the difference between a university and an LAC. More later.</p>
<p>OK, I got a little further. I still have the same question. Define “top”. What does that mean? Top to who? Top for what? Nothing has changed. It is meaningless because unless you define what you mean by top or best or whatever adjective you want to use, it means nothing. Since you seem unable to grasp this, I will indeed move on.</p>
<p>What I am curious about is why Emory fell again after it was at 9th… was the formula for calculating the rankings changed again?</p>
<p>fallenchemist :</p>
<p>Dont panic. Dont get frustrated. Its a new concept. Read on. Just focus on those responses posted by me and co-host tk21769. Upon completing reading it, if you still have the same doubts, we may figure out the issues that caused you falter, together. Using offensive languages like those does not help you understand the beauty of modeling.</p>